Card et al. constructed a design space for the generation and test of input devices designs.
For generating input devices designs, they built on Foley et al. and Buxton et al.’s work where they applied Mackinlay’s Semantic theories and considered the language of input device interaction (primitive movement vocabulary and composition operators) transiting across three agents (human, user dialogue machine, and application).
For testing input devices designs, they consider two figures of merit: expressiveness (the input conveys exactly and only the intended meaning) and effectiveness (the input conveys the intended meaning with felicity).
* Previous work on human-machine input devices has provided three lines of development in this area: toolkits, taxonomies, and performance studies;
* In order to achieve a systematic framework for input devices, toolkits need to be supported by technical abstractions about the user, the devices themselves, and the task they are used in performing.
* An input device is part of the means used to engage in dialogue with a computer or other machine.
* Semantic theories provide the means by which the design space can be generated. Human performance theories provide the means by which design points in the space can be tested.
* In this paper, we built on the results from the earlier paper and proceed to the use of human performance theories and data for the evaluation of points in this design space.
* Basically, an input device is a transducer from the physical properties of the world into logical values of an application (Buxton & Baeker).
* Three composition operatiors:
** Merge composition is the combining of two devices such that the resulting input domain set is the cross product of the input domains of the two devices;
** Layout composition is the collocation of two devices on different places of a common panel;
** Connect composition occurs when the output domain of one device is mapped onto the input domain of another device.
# Consider the difference between input measures and input resolutions.
# Is it time to do the same thing for ubiquitous computing input devices? (first, what is that? we have not really got mature *ubicomp* input devices – most stay at research levels)
# wording: three lines of development
# wording: sketch only a limited picture of the design space of input devices and their properties